sign in sign up
<<
Home , Computer mouse, Gamepad, Input device, Pointing device

EXAMENSARBETE INOM TEKNIK, GRUNDNIVÅ, 15 HP STOCKHOLM, SVERIGE 2018

Identifying effectiveness of different devices as pointing devices for graphical user interfaces

TED KLEIN BERGMAN

RONYA BUDAK

KTH SKOLAN FÖR ELEKTROTEKNIK OCH DATAVETENSKAP

Identifying effectiveness of different input devices as pointing devices for graphical user interfaces

Ronya Budak Ted Klein Bergman KTH, Royal Institute of , School of KTH, Royal Institute of Technology, School of and Science, Electrical Engineering and Computer Science, Department of Media Technology and Interaction Department of Media Technology and Interaction Design Design [email protected] [email protected]

ABSTRACT handhållna enheten. Resultatet tyder på att ritplattan kan vara en As new input devices are introduced to the technological market effektiv pekdon i händerna på en erfaren användare än bevisat här, and an increasing part of the population are familiar with the men för nya användare kan det inte fungera lika bra som usage of digital devices, it may seem difficult to know which datormusen överlag. input devices to aim for when developing new programs and Keywords applications that use graphical interfaces. According to a previous study in this area, the was the most preferred and ; computer mouse; handheld controller; performed best when tested in speed and accuracy when tablet; comparison; graphical ; . compared to the keyboard and drawing tablet even though the 1. INTRODUCTION keyboard was the device most were familiar with. This paper Graphical user interfaces for computer applications and are attempted to recreate the study with the exception of testing more mainly focused at providing the user an interface to interact with things, such as number of miss-clicks, distance travelled, graphical components [8]. Doing this usually requires navigating directional changes and time. The keyboard was replaced by a a on a two-dimensional screen and creating some sort of an handheld controller as well. The results were that the mouse (i.e. clicking/touching), often referred to as a pointing performed best yet again, and proved to be easy to use efficiently device. It’s still possible to navigate with other input devices, such for both new and experienced users. The handheld controller was as a keyboard, but this has been shown to be slow [2, 5]. shown not to be an optimal pointing device, mostly due to it being stuck to a fixed speed and not able to accelerate. It did well in the Today the most used input device for graphical applications is the matter of amount off miss-clicks and directional changes, which computer mouse, but with new technological advances being can be tied to its slow speed. The drawing tablet was well introduced everyday it might be time to consider if another device received by the new users and experienced as more accurate than may work better for these cases. the handheld device. Its results suggest that the drawing tablet could be an efficient pointing device than proven here in the One commonly used device is the handheld controller, that many hands of an experienced user, but for new users it could not prefer to use within computer gaming, but not so much elsewhere. perform as well as the mouse device overall. The less used drawing tablet is mostly used for photo editing and SAMMANFATTNING digital drawing. Are these limited only to their current uses or Eftersom nya inmatningsenheter introduceras på den tekniska could they be better integrated in more common usages? This marknaden och en större del av befolkningen bekantar sig med paper aims to test three input devices, the computer mouse, the användningen av digitala enheter, kan det verka svårt att veta handheld controller and the drawing tablet, to how well they vilka inmatningsenheter man ska rikta sig till när man utvecklar compare against each other in common tasks for graphical user nya program och applikationer som använder grafiskt gränssnitt. interfaces. Enligt en tidigare studie var datormusen den mest föredragna 1.1 Problem definition and purpose enheten och fick bäst resultat inom snabbhet och noggrannhet, jämfört med tangentbordet och ritplattan inom detta område, trots 1.1.1 Research question att tangentbordet var den enhet som de flesta var bekanta med. What is the effectiveness of mouse, handheld controller and Denna undersökning försökte återskapa studien med undantaget drawing tablet within the areas of precision, speed and error att testa flera egenskaper, såsom antal missade klickar, avstånd, proneness in basic graphical user interfaces? byten av färdriktning och tid samt att tangentbordet ersattes av en handhållen kontroller. Resultaten var att musen presterade bäst än 1.1.2 Purpose en gång och visade sig vara lätt att använda effektivt för både nya An earlier article from 1994 by Coll, Richard [3], answered this och erfarna användare. Den handhållna styrenheten visade sig inte question for similar input devices. However, this was done 24 vara en optimal pekdon, främst på grund av att den hade en years ago, and technology and its integration in society has bestämd hastighet och inte kunde accelerera. Det gick bra i fråga changed much since then. Not only are people more affiliated om mängden missade klickar och riktningsförändringar, som kan with this type of technology today, but everyday technology has vara knutna till dess långsamma hastigheten. Ritplattan mottogs developed with an accelerating speed and there are more choices väl av de nya användarna och upplevdes som mer exakt än den of input systems today as well. This research aims to confirm whether their results still holds and to provide useful data which compared to each other in the sense of speed, accuracy, and developers can use when designing graphical user interfaces for preference [3]. one of the specified input devices, and for developers to determine That paper may not be as viable today due to the changes done which input device to focus their application for. within the fields of these devices and due to their integration into modern society. The study conducted in 1994 does not specify the 1.1.2 Delimitation specification of the used input devices, so it’s reasonable to Some changes to the study were made, with the most noticeable assume they used devices that have seen significant changes and one being that the keyboard was not tested and was instead are evidently not an appropriate indication of the modern versions. replaced by the handheld controller. This was due to the keyboard It was believed that the pen would prove to perform better than being significantly slower to use than both mouse and drawing the mouse and keyboard, however it was found that the mouse had tablet, which seems consistent in other research as well [2, 3, 5]. performed best out of the three devices used, despite that only 3 This study does not consider differences in sex either, as this has out of 63 users had used a mouse before. The mouse was also been shown to have no significant difference [3, 6]. most preferred by the testers for general work, but the keyboard This study chose to focus on hand-steered mechanical input was preferred for highly accurate work and the device all the devices, and thus disregard devices such as , participants were familiar with. eye-tracking, foot controlled devices and others. This was to both Newer studies have been carried out, although these mainly focus compare devices within a common category, but also to reduce on ergonomy/health [7], ease of use [12], or preference the number of devices to test, in order to cut down time in order to perception. Those who focus on the performance of a pointing allow the experiments to not take too long time. device tends to focus on newer technology, such as touch screens. It seems relevant to perform a study focusing on more current devices used today and how effectively they can be used as a 2. BACKGROUND pointing device. In this section the definition and history of the tested input devices In [12], 8 input devices were compared when users played a one will be given, as well as a short description of the experiment versus one in a 3D-environment, amongst which were a conducted by Coll, Zia and Coll in 1994 [3] and other previous computer mouse and an . The mouse and XBOX studies. controller were rated together with 2 other devices easiest to use, and XBOX gamepad was proven one of the most well adapt to 2.1 Computer mouse the test environment. The computer mouse is the first commercially successful analog pointing device. It was invented by of Stanford A study in 2016 [14] evaluated the mouse against handheld Research Center in 1964, and was patented as the X-Y Position controller (among others) both as a pointing device and in the ​ Indicator [13], before getting the more commonly known name perspective of usability and user experience. They concluded that Mouse or Computer Mouse due to the cord looking like the tail of the mouse is still the better of those two as a pointing device and ​ a mouse. [1] in the sense of usability and user experience. The controller performed the worst as a pointing device but was more likeable The mouse has seen many changes since it was originally for game play. introduced and today it is mostly recognized that a mouse has a right and a left and a wheel. Earlier it was From a study done 2000 [11], a Mouse, digitising pen, accupoint common to have a mechanical wheel beneath the mouse to detect and were tested against each other as pointing devices in motion, which in modern time been replaced by a laser sensor [9]. a 2D space, and results showed that the mouse and digitising pen did excellent as pointing devices and received similar results. 2.2 Handheld USB controller The handheld USB controller is a controller that plugs into an 2.5 Concepts USB port. It is a form of gamepad, a type of held Input device refers to one of the three devices: mouse, handheld in two hands, where the thumbs are used to provide input, used controller or drawing tablet. mostly for console and PC-games and often has two and Effectiveness refers to how well the devices perform in terms of several buttons. Its movement is based on one or two fixed speed and accuracy. speeds, and cannot accelerate. 3. METHOD 2.3 Drawing tablet Three different experiments were designed based on those done in The drawing tablet is an input device that allows for drawing and a previous study [3], each to simulate different actions, and they sketches to be directly entered into a computer. It is made up of an all were repeated three times each (three “rounds”). The subject electronic tablet paired together with a pen. The pen is similar to a tried each input device for 3 rounds for all three experiments. The normal ballpoint pen, but uses a head that can be detected by the order of which device each subject started and ended with was tablet that translate the movements into digital signals. The tablet randomized to prevent order effects, such as practice effect and represents the computer screen in a fixed manner with each point fatigue effect [10]. All experiments started with the cursor on the tablet having a fixed point on the screen, differing to how a centered at the screen, and the timing started when the subject computer mouse is used. [15] moved the cursor. A countdown of 2 seconds was added before each experiment to prevent the subject to accidently move the 2.4 Previous studies mouse and start the timer. General observations were made on the In a study done by Coll, Zia, Coll (1994), a computer mouse, participants performances with the devices. drawing tablet (“pen-on-vertical-screen”) and a keyboard were

After the experiments each participant were sent an online survey 3.3 The experiments to their mail where they were to rate their previous experience with each input device and which device they prefered for the Each experiment tested different areas relating to common experiment. They also were to comment for each device what it activities within graphical user interfaces, namely pointing, had done well and what it had done poorly. clicking, and dragging. In the first test, clicking and moving was tested to emulate 3.1 Equipment selecting. Each subject used the input device to control a cursor Three input devices were used. G502 Proteus Spectrum that they were to place over a square and click. Once the square Gaming Mouse (the mouse), HUION H610 Pro Drawing Tablet was clicked, it would disappear and a new one appeared at another (the drawing tablet) and USB-208 USB 1.0/2.0 Wired Game location. In total each subject pressed 10 squares at different Controller (the handheld controller). locations. The experiments began with the cursor positioned at the Each input device had a specific configuration that wasn’t center of the screen and started when they began to move the changed between the subjects. The mouse steered the cursor when cursor. moved across the table surface, and a click was created when the subject pressed left . All subjects used the right hand. The handheld controller steered the cursor with the help of its left , using his/her left thumb, and pressed with button number 3 on the right side, using his/her right thumb. The drawing tablet moved the cursor by hovering the pen above the tablet. To generate a click, the users touched the tablet with the tip of the pen. The experiments were created and run on a Macbook Pro 13 inch from early 2011. They were written in Python 3.6.2 using the external library Pyglet 1.3.1 for reading input and drawing graphics, and they ran at 1/60 second update rate. All experiments were carried out in a quiet, private room.

Figure 1: Experiment 1. Start screen. 3.2 The collected data ​ There were seven different attributes that were collected by the In the second test, dragging was tested to emulate drawing or program to be able to analyse the speed and accuracy. moving graphical components. The user was presented with ● Time - the average number of seconds it took each user several small squares laid out in the shape of a triangle, a to complete one round. rectangle and a pentagon. The subjects were to press on a square ● Miss-clicks - the average amount of clicks that wasn’t and hold the button as they moved over to another square. They ​ within the area expected from the experiment. then release the button in order to create a green line between each ● Precision deviation - the distance from the clickable square. For each completed shape, the current would disappear areas’ center. and the next shape would appear until the subject had drawn all ● Positions - a list of x, y-coordinates for each update three shapes. (1/60 second). Unfortunately, precision deviation weren’t recorded for experiment 1, and miss-clicks weren’t recorded for experiment 2. The positions attribute was recorded in order to calculate other ​ useful data. ● Distance - the number of the cursor traveled. Calculated by summing the distances between each recorded position. For diagonal movement, the pixels traveled would be √x2 + y2 , as according to pythagoras theorem. ● Speed - the average amount of pixels the cursor travels ​ (distance) divided by the time for each round. ● Direction changes - the number of times the user changed the direction either horizontally or vertically. Figure 2: Experiment 2. Experiment being executed. It’s number of horizontally direction changes is ​ calculated by counting how many times an x-coordinate In the last test, clicking and moving was tested to emulate picking minus the following x-coordinate changes signs. For up and dropping. The subjects had to move a shape into an outline vertical direction changes, the same was done using the of that shape. The shapes where lines in the form of a triangle, a y-coordinates. rectangle and a pentagon. They clicked once while the cursor was Direction changes, precision deviation and miss-clicks were put in over the shape to pick it up, moved it to the outline and clicked the category accuracy, whereas speed and time were grouped into the button again to “drop” the shape into the outline. Once a shape ​ ​ speed. had been placed correctly, a margin of error of 10 pixels between ​ the centres of the shapes, it would disappear and the next one 4. RESULTS would appear on screen. 4.2 Experiment results The results were analysed using one-way analysis of variance (ANOVA). The figures show graphs with results of the different aspects that were analysed, where the y-axis represents the measurements and the x-axis the experiment. Only results from experiments with significant results are shown. Figure 3 shows the average number of directional changes between each device for the three rounds. ANOVA showed this to be the most significant (F(2, 6) = 143.88, p < 0.00001, F(2, 6) = 136.83, p < 0.00001 and F(2, 6) = 203.68, p < 0.000001) difference of all attributes. This was the same through the other experiments as well. The scheffe test showed that the controller changed direction the least, followed by mouse and then tablet (controller < mouse < tablet) for all experiments. Figure 3: Experiment 3. Start screen. ​ Each subject were offered to test out the devices before the actual experiment started, in case they hadn’t tried the devices before, since this paper does not aim to analyze the learning curve of each device. The tests also had sound effects to indicate to the subject if they had done correctly or not. All three experiments (with three repetitions per experiment) all repeated for three rounds took about 25 minutes per subject. 3.5 Participants A total of 29 students between the ages of 18-30 (estimated) from KTH, Royal Institute of Science Stockholm School of Computer Science and participated in the study.

Figure 4: The average direction changes for experiments 1-3, for each ​ round. Direction changes represent the number of times the user changed the direction either horizontally or vertically. Significant differences were shown in all three experiments. Figure 5 shows the average speeds between each device for the three rounds. ANOVA showed this to be a significant difference (F(2, 6) = 100.42, p < 0.0001 and F(2, 6) = 145.36, p < 0.00001, F(2, 6) = 15.098, p < 0.01) for the first two experiments, with a lower significance for experiment 3. The scheffe test showed that mouse was the best (mouse > tablet > controller) for experiment 1 and 2, but for experiment 3 both mouse and drawing tablet where equally fast (mouse = tablet > controller).

Figure 5: Average speed for experiments 1-2. The speed is represented by ​ the average amount of pixels the cursor travels divided by time. 8. Significant differences were shown in all three experiments. As expected (since time and speed are related) the average time (seen in figure 6) between each device for the three rounds were shown to be significant (F(2, 6) = 27.249, p < 0.001, F(2, 6) = 572.851, p < 0.0001, and F(2, 6) =91.940, p < 0.0001) where experiment 1 had a lesser significance. The scheffe test showed that mouse and tablet took equal amount of time for all experiments, and controller the most time (controller > tablet = mouse).

Figure 8: Average miss-clicks for experiments 1 and 3 (whereas it was not recorded for the second experiment). Significant differences were shown in both experiments. Figure 9 shows the average distance. There were no significant differences (F(2, 6) = 0.2785, p < 1, F(2, 6) = 9.2335, p < 0.05, F(2, 6) = 0.1686, p < 1) between the devices for experiment 1 and 3 here. However, there was a significant between them in experiment 2, where the Scheffe test showed that both the tablet and controller required more distance than the mouse (tablet = Figure 6: Average time in seconds. The average number of seconds it ​ controller > mouse) to complete the task. took each user to complete experiments 1-3. Significant differences were shown in all three experiments. The tables presented here show the results of the most significantly differences recorded in each experiment. The first The average precision deviation (shown in figure 7) between each number represents which experiment the data is from, and the device for the two experiments (whereas the first experiment, it second from which round, ie: 1.2 references experiment 1, round wasn’t recorded) were shown to be significant (F(2, 6) = 17.443, p 2. < 0.01, F(2, 6) = 27.473, p < 0.001). The Scheffe test showed the mouse to be the most accurate, with drawing tablet and controller being equally accurate (controller = tablet > mouse) for all experiments.

Figure 9: Average distance, where distance is the number of pixels the ​ Figure 7: Average precision deviation for experiment 2 and 3 (whereas it cursor traveled. Significant differences were shown in all three ​ was not recorded for the first experiment). The precision deviation is the experiments. distance in pixels from the clickable areas’ ​center. Significant differences were shown in both experiments. 4.1 Survey results and observations The average miss-clicks (errors) between each device for the two 28 out of 29 participants answered the survey sent out to them. experiments (as for the second experiment, it wasn’t recorded) were shown to be significant (F(2, 6) = 20.756, p < 0.01, F(2, 6) = 24.818, p < 0.01). The mouse had the least amount of miss-clicks for experiment one, where drawing tablet and controller where equal (tablet = controller > mouse), and for experiment three both the mouse and controller performed better than the drawing tablet (tablet > controller = mouse). Results for this can be seen in figure

two comments on being steady (with the cursor), having smooth movements and doing well on the second and third test. Single comments were made on it clicking small targets, being fast, swapping directions, easy to use, easy to steer and it’s button being easy to press.

4.1.5 Controller - Negative comments 13 of the users complained about the speed of the device. 10 found it hard to move the cursor with the device when not moving in a straight line. 6 of the users didn’t like it’s precision and 2 comments were made on it being hard to hit the targets accurately, drawing lines and it being bad on the third test. Single comments were made on it being bad at the first test, all three tests and at hitting small objects. Figure 10: Testers familiarity with the devices where 1 stands for “rarely ​ use it” and 4 for “very experienced”. The y-axis represents the number of 4.1.6 Controller - Observations participants. It was observed that several testers had difficulty using this

device, and several complain about it’s speed during testing. This led to most of them adapting to moving the cursor in one direction while trying to stop or hit the button at the correct time when the cursor was over the desired location.

4.1.7 Drawing tablet - Positive comments 13 participants commented on the device being fast, and 10 thought it was precise. 8 users found it good at drawing lines and 4 thought it did well in the third test. 3 comments were made on it being intuitive and easy to use, while one comment was made on it being consistent, comfortable, easy to estimate the cursors movements, sensitive and on being fun to use.

4.1.8 Drawing tablet - Negative comments Figure 11: Most preferred devices for the experiments. 4 testers complained about it’s usage during the first test, while 3 ​ comments were made on it’s precision being bad and being unused to the device. 3 users also thought the hovering of the pen 4.1.1 Mouse - Positive comments to be difficult. 4 mentions were made on the pen reacting to shaky 14 participants found it’s precision was good and 11 that it had a hands and 2 on it being hard to understand the screen-tablet ratio good speed. 6 found that the mouse did well on everything, and 5 and having to move their hand around too much. Single comments commented on it being easy to control. 4 thought it did well on all were made on it being fast, difficult to estimate where the cursors the tests. 3 thought it did well in the first task and 2 that it did well position, not being ergonomic, sensitive and it being easy to in the third task. 2 mentioned being familiar with the device as a miss-click if the pen was tilted. positive trait. The devices click speed, accuracy, precise movements, small movements and less prone to miss-clicks got 4.1.9 Drawing tablet - Observations one positive comment. During testing it was observed that several of the users reacted positive initially to using the tablet, even though most of them had 4.1.2 Mouse - Negative comments no previous experience with it. It was observed that a few users 11 of the participants had nothing bad to say about the mouse. 7 would accidentally hold the pen at a much larger angle than users complained about the mouse speed and sensitivity but 3 of intended which would cause the cursor to go lower when they these pointed out that it might be because they weren’t used to the tried to click a target. When moving great distances with the pen it set speed. The device got 4 comments on not doing so well when was also observed that some testers would accidentally touch the drawing lines. 2 thought the mouse wasn’t as precise as the surface of the tablet with the pen several times. drawing tablet. One mention was made on the user having to move their hand and forearm too much. 5. DISCUSSION 5.1 Mouse 4.1.3 Mouse - Observations Nothing of value was observed during the experiments where the mouse was tested.

4.1.4 Controller - Positive comments 6 thought it had done nothing well. 5 said it was good at drawing straight lines. 3 thought it was ergonomic. The device received

The results from the tests suggest the mouse to be the best device perform best in any of the presented aspects and was the worst out of the three overall, only lacking in the area of directional amongst the three when it came to miss-clicks and directional changes, where the handheld controller did better. It also had to changes. change direction more often than the controller, which suggests that it overshoots (goes past the intended stopping/turning point It received a lot of positive comments in the survey, with almost inadvertently) its targets. half the group describing it as fast and over a third of them saying it was precise, even though it had the highest miss-click count of According to the survey results, the mouse was the most familiar the three devices. The comments section suggest this might be due and most preferred device to the test group (seen in figure 10), to some of the testers having shaky hands (and possibly not with 26/28 users stating that they were “very experienced” with it supporting their hands correctly on the surface), which should and 21/28 preferring it to the other two devices. 39% of the group explain it’s high number in directional changes as well. In the had nothing negative to comment on the mouse. 1/4th of the group observations it was noticed that when moving larger distances complained about the speed or sensitivity of the mouse, but 3 of with the pen, some users would accidentally touch the surface of these thought it might be due to it not being set to their personal the tablet with the tip of the pen. This is most likely due to preference and may in that case not be as relevant as otherwise inexperience with the device and may have caused it’s miss-click suggested. According to the positive comments, the mouse was count to be higher than it otherwise would have been. precise, had a good speed and was easy to use. This could be argued to be due to the testers familiarity with the device, but The ratio between people preferring a device and having used it according to [3], where only 3 out of 63 participants had used a before was the greatest for the drawing tablet, suggesting that mouse before, it still performed better than the keyboard, even many of the users would be open to use a drawing tablet. These though most of the participants had used the keyboard before. results suggest that the drawing device could be an efficient This strongly suggests the mouse device to be easy to use pointing device in the hands of an experienced user. For new users efficiently for both new and experienced users. Study [14] also it was well received but did not perform as well as the mouse confirms the mouse to be dominant in performance and device. In another study, [11] the mouse and digitizing pen were preferability. shown to have very similar results and thought to be great pointing devices, which supports the results of this study 5.2 Handheld controller somewhat. The handheld controller performed best in terms of directional 5.4 Method critique changes, which can also be seen in the survey as well due to it being said that the device had “smooth” movements. Otherwise, it There were a few bugs in the program created for the tests, which performed poorly in most aspects, speed being it’s greatest caused minor problems for about one third of the test group, hindrance. causing them to have to repeat the last step of some of the tests. This will have affected their time by adding to how long it took The only positive pointers from the survey were that the handheld them to finish the test. These may also have added slightly to the controller had smooth movements and was easy to use. About half click- misclick ratio and the distance moved due to the test being the group had trouble with it’s speed and and it was observed that longer than intended. There was one bug in the program that led the users had a hard time moving the cursor to the desired location to data not being recorded for the first few tests. This was on the screen and making finer adjustments, something all the counteracted when calculating the average, which was based on comments from the survey seem to confirm. the number of subjects the data was recorded from. Since it isn’t always a solid 29 subjects, this was taken into account when doing Over a third of the test group found it hard to be precise with this the each calculation for each mean and have been corrected. device, however data suggests that it performed equally as well as Nothing else was changed or corrupted by any means. the drawing tablet in precision deviation, and had fewer miss-clicks. This may be explained due to the drawing tablet being quite a bit faster than the handheld device, and was thus Mouse sensitivity is very personal, and thus the mouse controllers experienced as more accurate. This could also suggest that the speed was experienced differently among the testers. Only one handheld device may have only performed better due to being speed was programmed for the handheld controller as well, when more familiar to the test group than the drawing tablet. This can in most functions/games it has two, which made it badly adapted be compared to study [12] where it was shown that people to all the tests. Testers experienced it as too slow for the first test, familiar with the used input system of an XBOX controller had a and too fast for the second and third one. The quality between the small advantage over those were not familiar with their device of devices varied as well. The drawing tablet was a higher quality, choice. The results from figure 8 show that the handheld professional device, whereas the handheld controller was low controller did not receive any votes when the group was asked quality, hobby controller. The mouse was about average quality. which device they preferred, which together with the data and This will have affected speed and movement the most. The mouse comments made suggests that the handheld controller is very was used without a mouse mat since the table used was deemed inefficient as a pointing device. However, this does not prove it to appropriate to use without one, which may have affected mouse be a bad input device. According to [14], the controller performs precision slightly. badly at being a pointing device, but could have potential in a game environment, confirmed by [12] as well. 6. CONCLUSION

5.3 Drawing tablet As it stands, the mouse is still the most dominant input device. It is both faster and often more accurate, and most preferred which The drawing tablet performed equally as good as the mouse in is in line with study done by Coll [3]. Comparing to his study, a speed and time, performed equally as well as the controller on clear conclusion can be drawn that the mouse device could be precision deviation and in terms of travelled distance. It did not efficiently used by new and more experienced users. The handheld controller was proven not to be an optimal pointing doi:10.1016/s0020-7373(86)80034-7. device, due to it being stuck to a fixed speed, whereas the mouse and drawing tablet could accelerate the cursors speed when in use. [7] Keir, Peter J., et al. “Effects of Computer Mouse Design and It was however easier to minimize number of directional changes Task on Carpal Tunnel Pressure.” Ergonomics, vol. 42, no. 10, ​ ​ with the handheld controller device. The drawing tablet was the 1999, pp. 1350–1360., doi:10.1080/001401399184992. device users had least experience with, yet one fourth of the test group preferred it to the other devices. It was experienced as more [8] Levy, Steven. “.” Encyclopædia ​ accurate than the handheld device. The results suggest that the Britannica, Encyclopædia Britannica, Inc., 29 Mar. 2018, ​ drawing tablet could be an efficient pointing device in the hands www.britannica.com/technology/graphical-user-interface. ​ of an experienced user, but for new users it could not perform as well as the mouse device. This suggests that further research [9] Mouse.", "Computer. “Computer Mouse.” How Products Are ​ should be made with more experienced users of the drawing Made, Encyclopedia.com, 2018, ​ tablet, or a test group that learn to use it over a period of time to www.encyclopedia.com/manufacturing/news-wires-white-papers- further study the effectiveness of this device. and-books/computer-mouse. ​

7. FURTHER RESEARCH [10] “Order Effects.” Psychology, ​ ​ There are several things that could be extended to this study, such psychology.iresearchnet.com/social-psychology/social-psycholog as evaluating other characteristics between the input devices, y-research-methods/order-effects/. ​ including ergonomics, health, price, learning curve and more. This study also didn’t follow any accepted ISO standards (ISO 9241-9) [11] Phillips, James, and Thomas Triggs. “Cursor Control Device for testing computer pointing devices [4], which could be a Characteristics.” Australasian Journal of Information Systems, ​ ​ discredit to the study. Further research should be made on the vol. 7, no. 2, Jan. 2000, doi:10.3127/ajis.v7i2.269. effectiveness of the drawing tablet as a pointing device with users who are more experienced with the device or allowed to learn to [12] Thorpe, Andrew, et al. “History and Alternative Game Input use it over a period of time due to its promising results. Methods.” 2011 16th International Conference on Computer ​ Games (CGAMES), 2011, doi:10.1109/cgames.2011.6000321. ​ 8. ACKNOWLEDGMENTS Big thanks to Andre Holzapfel for his guidance throughout the [13] “US3541541A - X-y Position Indicator for a Display course of this study, and to the study group for their valuable System.” Google Patents, Google, ​ ​ input and suggestions. Special thanks to Josefin Ström as well for patents.google.com/patent/US3541541A/en. ​ helping us book rooms to perform the experiments in. [14] Young, Gareth, et al. “Usability Testing of 9. REFERENCES Controllers: A Case Study.” Games User Research, Mar. 2016, ​ ​ pp. 145–188., doi:10.1201/b21564-8. [1] Administrator, and MRW Connected. “Mouse.” A Lifetime ​ Pursuit - Doug Engelbart Institute, [15] “Digitizing Tablet.” The Five Generations of - ​ ​ www.dougengelbart.org/firsts/mouse.html. Webopedia Reference, ​ ​ www.webopedia.com/TERM/D/digitizing_tablet.html. ​ [2] Albert, Alan E. “The Effect of Graphic Input Devices on Performance in a Cursor Positioning Task.” Proceedings of the ​ Human Factors Society Annual Meeting, vol. 26, no. 1, 1982, pp. ​ 54–58., doi:10.1177/154193128202600114.

[3] Coll, Richard, et al. “A Comparison of Three Computer Cursor Control Devices: Pen on Horizontal Tablet, Mouse and Keyboard.” Information & Management, vol. 27, no. 6, ​ ​ 1994, pp. 329–339., doi:10.1016/0378-7206(94)90014-0.

[4] Douglas, Sarah A., et al. “Testing Pointing Device Performance and User Assessment with the ISO 9241, Part 9 Standard.” Proceedings of the SIGCHI Conference on Human ​ Factors in Computing Systems the CHI Is the Limit - CHI '99, ​ 1999, doi:10.1145/302979.303042.

[5] Goodwin, Nancy C. “Cursor Positioning on an Electronic Display Using Lightpen, Lightgun, or Keyboard for Three Basic Tasks.” Human Factors: The Journal of the Human Factors and ​ Ergonomics Society, vol. 17, no. 3, 1975, pp. 289–295., ​ doi:10.1177/001872087501700308.

[6] Karat, John, et al. “A Comparison of Selection Techniques: Touch Panel, Mouse and Keyboard.” International ​ Journal of Man- Studies, vol. 25, no. 1, 1986, pp. 73–88., ​

APPENDIX A - The experiments ​ Experiment 1

Experiment 2

*

Experiment 3

www.kth.se